Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2015-02-05

| Channels | #marpa index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:11 jeffreykegler As most of you probably know, Larry has set a tentative date for Perl 6.0.0, including a list of things to complete before release.
00:12 jeffreykegler It has 3 major items, one of which is from the Marpa point of view, an anti-feature -- it makes a Perl 6 Marpa less likely, rather than more.
00:12 jeffreykegler This feature is NFG -- a new approach to Unicode.
00:13 jeffreykegler Marpa's problem is that AFAICT, it makes the translation from character to bytestream and back, very complex, and all of that complexity must be dealt with by any C language module inside Perl 6.
00:14 jeffreykegler This work would be IMHO, in a worthless cause.
00:15 jeffreykegler Unicode, in its design IMHO, crossed the line from representing streams of the basic glyphs of a language, over into representing matters of typography.
00:16 jeffreykegler That is, there is a line at which an excellent, fully-feature system for representing glyphs, turns itself into an underpowered encoding of typography.
00:18 jeffreykegler An example (I will stick to my native language, which fortunately is the lingua franca of computing) is the grapheme for the ligature 'ffl'.  Unicode IMHO should never have bothered with this, and just let 'ffl' be f then f then l, leaving the ligatures for a typographer to figure out.
00:19 jeffreykegler NFG tries to go beyond the Unicode committee in this matter, when they've already gone too far.
00:19 jeffreykegler Anyway, bottom line, if Perl 6.0.0 came out without NFG, I would be happier rather than sadder.
00:19 jeffreykegler Especially if that meant it came out faster.
00:38 koo5 'LATIN SMALL LIGATURE FFL' sounds like one of those round-trip compatibility cases
00:38 koo5 "Unicode is a messy thing. You wouldn't think that something so trivial as representing all the text in all the languages in all the world, keeping track of the relationships between various related symbols and characters, and creating a clear delineation between languages and dialects would be such a hassle, but apparently it is"
00:38 koo5 hehe
01:52 ronsavage joined #marpa
02:49 ilbot3 joined #marpa
02:49 Topic for #marpa is now Start here: http://savage.net.au/Marpa.html - Pastebin: http://scsys.co.uk:8002/marpa - Jeffrey's Marpa site: http://jeffreykegler.github.io/Marpa-web-site/ - IRC log: http://irclog.perlgeek.de/marpa/today
05:12 jdurand joined #marpa
05:16 jdurand Re http://irclog.perlgeek.de/marpa/2015-02-05#i_10062637 - this could be a revival for the need of marpaWrapper, quite finished but not yet, that takes the approach to have a "reader" - i.e. it is asking the user what is the next codepoint instead of doing the read himself - for the eventual marpaXml, I'll still rely on ICU which I believe or hope will suppport NFG
05:29 ronsavage joined #marpa
06:18 ronsavage joined #marpa
09:17 lwa joined #marpa
15:03 koo5 joined #marpa
21:44 ronsavage joined #marpa
21:57 jeffreykegler joined #marpa
22:06 ronsavage Just starting to analyze test failures with latest Marpa::R2.
22:15 ronsavage One module's problems fixed. I changed the output indentation of Tree::DAG_Node, and am now finding the modules which had the output tree hard-coded in their tests.
22:15 ronsavage One module still to be examined.
22:27 ronsavage Phew: Other module fixed too.
22:29 ronsavage The modules are MarpaX :: Demo :: StringParser and MarpaX :: Grammar :: Parser.
22:30 ronsavage Marpa::R2 V 2.103004. Test statistics:
22:30 ronsavage Fails: 0. Files: 41. Modules: 10. Passes: 10. Tests: 419. Total: 419.
22:30 ronsavage Duration: 54 seconds
22:42 ronsavage New versions are now on CPAN.
22:43 jeffreykegler ronsavage: great news, then.
22:44 ronsavage The # of tests has finally gone over 400 because of the 2 new modules Text::Balanced::Marpa and Text::Delimited::Marpa.
22:44 ronsavage jk: Marvellous news, and a relief for both of us :-)/
22:44 jeffreykegler I know your Marpa style is event-intensive, and I hacked the event code a lot, so I was wondering how this would go.
22:45 ronsavage Sure, and I didn't know what to expect, but your testing is pretty thorough, so I did somewhat suspect my code.
22:47 jeffreykegler I have considered cutting back on my test suite, which is very large by CPAN standards, but I've decided the large test suite is the better of the choice of possible evils.
23:06 ronsavage I vote against cutting back your test suite. After all, you can't expect /us/ to catch all /your/ mistakes. Hahahaha.

| Channels | #marpa index | Today | | Search | Google Search | Plain-Text | summary